[styayur]

Summary

Stability patch based on the published v1.1.0 release.

• Matched baseline entries created under an older rule-semantics version now become Stale and require human review instead of remaining silently accepted.
• Benchmark environment metadata now uses the benchmark project root for Git revision/worktree state rather than CodeForge's current working directory.
• Clean benchmark worktrees are now recorded explicitly.
• Versions and release notes are aligned to 1.1.1.

Why this is real

Auditing v1.1.0 found:

1. assessment_from_entry() always returned Active, even when a matched entry's rule version was old.
2. Benchmark environment discovery used the process cwd, so running codeforge benchmark /path/to/project could record CodeForge's commit instead of the target project's commit.

Compatibility

• Baseline schema remains 1.
• Existing baseline files remain readable; old-rule matches become human-review items.
• Review and benchmark JSON changes are additive/correctional.
• SARIF and Fleet formats are unchanged.
• MSRV remains Rust 1.90.

Verification

• Rust fmt, Clippy, workspace tests and release CLI build passed.
• Added matched-stale baseline regression test.
• Added target-root benchmark environment test.
• Manual benchmark from another cwd recorded the target repository's commit and clean worktree state.
• Frontend lint/typecheck/build and Markdown checks passed.